eng_spa_seq2seq

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0656

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2281	0.032	500	0.2171
0.194	0.064	1000	0.1832
0.1684	0.096	1500	0.1612
0.1583	0.128	2000	0.1476
0.1451	0.16	2500	0.1344
0.1371	0.192	3000	0.1238
0.1286	0.224	3500	0.1164
0.1231	0.256	4000	0.1099
0.1191	0.288	4500	0.1048
0.1119	0.32	5000	0.0997
0.1072	0.352	5500	0.0956
0.1073	0.384	6000	0.0917
0.0961	0.416	6500	0.0887
0.0983	0.448	7000	0.0865
0.0942	0.48	7500	0.0834
0.0921	0.512	8000	0.0814
0.0901	0.544	8500	0.0792
0.0853	0.576	9000	0.0771
0.0846	0.608	9500	0.0761
0.0823	0.64	10000	0.0739
0.0823	0.672	10500	0.0727
0.0824	0.704	11000	0.0717
0.081	0.736	11500	0.0709
0.079	0.768	12000	0.0695
0.0777	0.8	12500	0.0686
0.0759	0.832	13000	0.0676
0.0769	0.864	13500	0.0672
0.0781	0.896	14000	0.0666
0.0747	0.928	14500	0.0662
0.0757	0.96	15000	0.0658
0.0783	0.992	15500	0.0656

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Tokenizers 0.20.3